This competing renewal application seeks to extend, for a further 5 years, the Monongahela-Youghiogheny Healthy Aging Team (MYHAT) project, a 10-year longitudinal study of a representative population-based cohort in an economically depressed small-town region near Pittsburgh, PA. MYHAT has thus far focused on characterizing MCI and identifying risk factors for progression from MCI to incident dementia, and identified considerable heterogeneity in risk relationships. Building on findings to date, our new aims are: (1) To identify risk /protective factors for dementia in those aged 90+ years, in whom incidence continues to rise but few risk factors have been found; (2) To identify vascular / inflammatory/ metabolic risk factors, particularly diabetes-related factors, for dementia in the cohort as a whole ; (3) To examine birth cohort effects in the incidence of dementia and cognitive decline, and determine whether they can be explained by secular trends in risk and protective factors; (4) To identify genes associated with cognitive decline endophenotypes using existing genome-wide genotype data and then confirming in a replication sample. We also propose an exploratory aim to extend amyloid and tau PET neuroimaging into a subgroup of 120 MYHAT participants, at no cost to this study, as part of an independent project. Here the scientific objective is to identify factors associated with cerebral amyloid and tau deposition and determine whether those variables predict dementia in the MYHAT cohort as a whole. The original MYHAT cohort (N~2,000) aged 65+ years at study entry is now aged 75+ and has 10 years of rich longitudinal data that are ideal for achieving our aims. About half the cohort has been lost to attrition, including illness and mortality, over the past 9 years. We will replenish te cohort by recruiting 700 new participants aged 65-74 who are able to contribute the maximum person-years of followup, and will also enhance our ability to investigate cohort effects. Our over-arching theme is to investigate the heterogeneity of MCI and dementia in the population at large, deconstructing the outcomes as well as the predictor variables to identify finer-grained relationships which will shed light on underlying disease mechanisms. This approach is consistent with the goals of translational epidemiology, going beyond simple description to identify risk and protective factors with clinical and public health significance, and the subgroup in which these factors operate; generate and test mechanistic hypotheses; and identify population trends with implications for policy and planning.